Sex Differences in Cognitive and Symptom Profiles of Older Adults with Autism Spectrum Disorder

Background: Theories have suggested that females with autism spectrum disorder (ASD) demonstrate higher levels of symptom masking relative to males, and a proposed masking mechanism is executive functioning. Declines in executive functioning are a hallmark of normal aging and are also observed in aging adults with ASD, suggesting that aging may reduce masking abilities. Furthermore, if females with ASD are preferentially engaging executive function strategies for symptom masking, aging may have a greater effect on age-related symptom exacerbation for females than males. However, no studies have characterized executive function sex differences in an older adult sample and its relationship with ASD symptoms.

Objectives: We aimed to provide a preliminary characterization of sex differences on two measures of executive functioning in mid-to-older adult men and women with ASD and its association with ASD traits.

Methods: Participants were mid-to-older (40-70 years) adult men or women of average intellectual functioning with ASD or NT development (female ASD: n=11; female NT: n=12; male ASD: n=29; male NT: n=23). Sex, diagnosis, and sex by diagnosis interactions were examined using Analysis of Variance for a behavioral regulation (Wisconsin Card Sorting Task; WCST) and a metacognitive (Tower of London; ToL) measure of executive functioning. Relationships between executive functioning and ASD traits were examined using the Social Responsiveness Scale – 2^nd Edition (SRS-2) Social Cognition Subscale and the Adult Repetitive Behaviors Scale (RBQ-2A). IQ was included as a covariate in all analyses.

Results: For WCST, there was a significant main effect of diagnosis (F_1,71=12.878, p=0.001) and a significant diagnosis by sex interaction (F_1,71=4.030, p=0.049). Both women and men with ASD made more errors than their NT counterpart (women: p=0.004, novel finding; men: p=0.044, one-tailed replication); but, the magnitude of difference was greater for women with ASD (women: d=0.97; men: d=0.48; Fig. 1a). No significant main effects or interaction were observed for the ToL. However, exploratory post hoc analysis showed that females with ASD demonstrated more planning errors than males with a moderate effect size (d=0.46) that approached significance (p=0.107; Fig. 1b). In male and female ASD groups, correlations were observed between the SRS-2 Social Cognition Subscale and ToL performance, such that worse ToL performance was associated with increased ASD-related social behavior (female ASD: r₇=0.70, p=0.036; male ASD: r₂₇=0.39, p=0.045; Fig. 2a). Alternatively, the RBQ-2A showed a correlation with WCST performance in females with ASD but not males (female ASD: r₈=0.70, p=0.026; male ASD: r₂₁=-0.09, p=0.69), such that more errors were associated with increased repetitive behaviors (Fig. 2b).

Conclusions: In one of the first investigations of sex differences in mid-to-older adults with ASD, this study shows a tendency toward greater executive function difficulties in women with ASD, compared to men with ASD. Furthermore, women with ASD showed stronger associations between ASD traits and executive functioning than men. This suggests that older adult women with ASD may capitalize even further on executive functioning for symptom masking relative to men with ASD. Longitudinal research on sex differences in age-related cognitive and symptom changes are warranted.